Upload folder using huggingface_hub

db.py

@@ -1,3 +1,4 @@
+import time
 import pandas as pd
 import numpy as np
 import random
@@ -20,8 +21,21 @@ logger.setLevel(logging.INFO)
 
 
 class SanatanDatabase:
-    def __init__(self) -> None:
+    _instance = None
+
+    def __new__(cls, *args, **kwargs):
+        # ✅ Ensure only one instance exists
+        if cls._instance is None:
+            cls._instance = super().__new__(cls)
+            cls._instance._init_once()
+        return cls._instance
+
+    def _init_once(self):
+        """Initialize once per process"""
         self.chroma_client = chromadb.PersistentClient(path=SanatanConfig.dbStorePath)
+        self._count_cache = {}  # {collection_name: (timestamp, count)}
+        self._cache_ttl = 84600   # seconds (24 hours)        
+        logger.info("✅ SanatanDatabase singleton initialized")
 
     def does_data_exist(self, collection_name: str) -> bool:
         collection = self.chroma_client.get_or_create_collection(name=collection_name)
@@ -540,11 +554,26 @@ class SanatanDatabase:
         )
 
     def count(self, collection_name: str):
+        # check cache first
+        now = time.time()
+        cached_entry = self._count_cache.get(collection_name)
+        if cached_entry:
+            ts, cached_count = cached_entry
+            if now - ts < self._cache_ttl:
+                logger.debug("Cache hit for collection [%s]: %d", collection_name, cached_count)
+                return cached_count
+            else:
+                logger.debug("Cache expired for [%s]", collection_name)
+
+        # fetch fresh count
         collection = self.chroma_client.get_or_create_collection(name=collection_name)
         total_count = collection.count()
         logger.info("Total records in [%s] = %d", collection_name, total_count)
-        return total_count
 
+        # update cache
+        self._count_cache[collection_name] = (now, total_count)
+        return total_count
+        
     def test_sanity(self):
         for scripture in SanatanConfig().scriptures:
             count = self.count(collection_name=scripture["collection_name"])

server.py

@@ -1,12 +1,12 @@
 # server.py
 import random
 import traceback
-from typing import  Optional
+from typing import Optional
 import uuid
 from fastapi import APIRouter, HTTPException, Request, Query
 from fastapi.responses import JSONResponse
 import pycountry
-from pydantic import BaseModel
+from pydantic import BaseModel, Field
 from chat_utils import chat
 from config import SanatanConfig
 from db import SanatanDatabase
@@ -119,13 +119,13 @@ async def handle_fetch_languages():
     languages.sort(key=lambda x: x["name"])
     return languages
 
+
 @router.get("/languages_v2")
 async def fn_handle_fetch_languages_v2():
     val = await handle_fetch_languages_v2()
     return val
 
 
-
 @router.post("/greet")
 async def handle_greet(msg: Message):
     markdown = "Namaskaram 🙏 I am **bhashyam.ai** and I can help you explore the following scriptures:\n---\n"
@@ -212,6 +212,9 @@ async def handle_get_scriptures():
 class ScriptureRequest(BaseModel):
     scripture_name: str
     unit_index: int
+    request_id: str | None = Field(
+        default=None, alias="request_id"
+    )  # optional, backward compatible
 
 
 @router.post("/scripture")
@@ -219,26 +222,35 @@ async def get_scripture(req: ScriptureRequest):
     """
     Return a scripture unit (page or verse, based on config),
     including all metadata fields separately.
-    used for page view to fetch by global index.
+    Used for page view to fetch by global index.
     """
-    logger.info("get_scripture: received request to fetch scripture: %s", req)
+    # ensure we have a valid request id (deviceId)
+    request_id = req.request_id or f"auto-{uuid.uuid4()}"
+    logger.info(
+        f"get_scripture: received requestId={request_id}, scripture={req.scripture_name}, unit_index={req.unit_index}"
+    )
 
     # find config entry for the scripture
     config = next(
         (s for s in SanatanConfig().scriptures if s["name"] == req.scripture_name), None
     )
     if not config:
-        return {"error": f"Scripture '{req.scripture_name}' not found"}
+        return {
+            "error": f"Scripture '{req.scripture_name}' not found",
+            "requestId": request_id,
+        }
 
     # fetch the raw document from DB
     raw_doc = SanatanDatabase().fetch_document_by_index(
         collection_name=config["collection_name"],
         index=req.unit_index,
-        # unit_name=config.get("unit_field", config.get("unit")),
     )
 
     if not raw_doc or isinstance(raw_doc, str) or "error" in raw_doc:
-        return {"error": f"No data available for unit {req.unit_index}"}
+        return {
+            "error": f"No data available for unit {req.unit_index}",
+            "requestId": request_id,
+        }
 
     # canonicalize it
     canonical_doc = SanatanConfig().canonicalize_document(
@@ -249,8 +261,8 @@ async def get_scripture(req: ScriptureRequest):
 
     # add unit index & total units (so Flutter can paginate)
     canonical_doc["total"] = SanatanDatabase().count(config["collection_name"])
+    canonical_doc["requestId"] = request_id
 
-    # print("canonical_doc = ", canonical_doc)
     return canonical_doc
 
 
@@ -277,7 +289,7 @@ async def get_scripture_configs():
             {
                 "name": s["name"],  # e.g. "bhagavad_gita"
                 "title": s["title"],  # e.g. "Bhagavad Gita"
-                "banner_url": s.get("banner_url",None), 
+                "banner_url": s.get("banner_url", None),
                 "category": s["category"],  # e.g. "Philosophy"
                 "unit": s["unit"],  # e.g. "verse" or "page"
                 "unit_field": s.get("unit_field", s.get("unit")),
@@ -335,12 +347,12 @@ async def search_scripture_find_first_match(
                 page_size=None,
             )
         else:
-             # optimization. get only first match if no has_audio parameter is provided.
-             result = db.fetch_first_match(
+            # optimization. get only first match if no has_audio parameter is provided.
+            result = db.fetch_first_match(
                 collection_name=config["collection_name"],
                 metadata_where_clause=filter_obj,
             )
-             results = {
+            results = {
                 "ids": list(result["ids"]),
                 "documents": list(result["documents"]),
                 "metadatas": list(result["metadatas"]),
@@ -578,10 +590,11 @@ def route_get_donation_product_ids(include_tests: bool = False):
         ]
     return products
 
+
 @router.get("/discourse/list")
 async def get_all_discourses(
     page: int = Query(1, ge=1, description="Page number (1-indexed)"),
-    per_page: int = Query(10, ge=1, le=100, description="Number of items per page")
+    per_page: int = Query(10, ge=1, le=100, description="Number of items per page"),
 ):
     """
     Returns a paginated list of discourse topics.
@@ -593,6 +606,7 @@ async def get_all_discourses(
     result = await get_discourse_summaries(page=page, per_page=per_page)
     return result
 
+
 @router.get("/discourse/find/{topic_id}")
 async def get_discourse_detail(topic_id: int):
     """
@@ -601,4 +615,4 @@ async def get_discourse_detail(topic_id: int):
     topic = await get_discourse_by_id(topic_id)
     if not topic:
         raise HTTPException(status_code=404, detail="Discourse topic not found")
-    return topic
+    return topic